CN101588341B - Lost frame hiding method and device thereof - Google Patents

Abstract

The invention discloses a method for concealing lost frames, which includes: when detecting the loss of MDCT coefficients, using the historical signals before the lost frames corresponding to the MDCT coefficients to generate a first composite signal; Perform fast IMDCT transformation on a synthesized signal to obtain IMDCT coefficients corresponding to the lost MDCT coefficients; use the IMDCT coefficients corresponding to the lost MDCT coefficients and the IMDCT coefficients adjacent to the IMDCT coefficients to perform TDAC processing to obtain the The signal corresponding to the lost frame. Correspondingly, the invention also discloses a device for realizing frame loss concealment. A frame loss concealment method and device implementing the present invention recover high-quality voice signals and improve QoS by making full use of received partial signals.

Description

Method and device for hiding frame loss

technical field

The invention relates to the field of electronic communication, in particular to a method and device for concealing frame loss.

Background technique

With the development of network technology, the application of real-time voice communication through packet switching network to transmit voice packets is becoming more and more extensive, such as IP-based voice communication (Voice Over IP, VoIP). However, networks based on packet-switching technology were not originally designed for applications requiring real-time communication, and are not inherently infallible. During the transmission process, the data packet may be lost, or it may be delayed and discarded by the receiving end after the time point of playback. In the above two cases, it is considered that a network packet loss has occurred. Network packet loss is a big problem for the real-time and voice quality required by VoIP. The receiving end of VoIP is responsible for decoding the voice packets sent by the sending end into playable voice signals, but in the case of packet loss, if the lost voice packets are not compensated, the voice signal will become discontinuous and Noisy signal is generated, which affects the quality of voice calls. Therefore, a robust packet loss concealment method is needed in real-time communication systems to recover lost data packets, so that good call quality can still be obtained in the case of network packet loss.

At present, the most commonly used packet loss concealment technology is the packet loss concealment technology based on pitch repetition. For example, a set of voice compression standard G.711 Appendix I formulated by the International Telecommunication Union adopts a packet loss concealment method based on pitch waveform substitution. Pitch waveform replacement is a processing technology based on the receiving end to compensate for lost audio frames. It first uses the historical signal before the lost frame to calculate the pitch period T ₀ of the historical signal, and then repeats the length of T ₀ before the lost frame. A segment of the signal to reconstruct the signal corresponding to the lost frame. As shown in Figure 1, frame 2 is a lost frame with a frame length of N, and frame 1 and frame 3 are both good frames. Assuming that the pitch period corresponding to the historical signal (the signal before frame 1 (including frame 1)) has been obtained is T ₀ , and the corresponding interval of the signal is interval 1; then the signal corresponding to the last pitch period of the historical signal (i.e. The signal corresponding to interval 1) is repeatedly copied to frame 2 until it is full, so as to reconstruct the signal corresponding to the lost frame. In Fig. 1, the signal of two pitch periods needs to be copied repeatedly to fill up the lost frame.

However, if the signal that repeats the last pitch in the historical signal is directly used as the signal corresponding to the lost frame, then a waveform mutation will occur at the junction of the two pitches. In order to ensure the smoothness of the splicing, usually before the lost frame is filled with the signal of the last pitch period in the history buffer, cross-fading is performed on the last T ₀ /4 signal of the history buffer. As shown in Figure 2, the window used is a simple triangular window, where the rising window corresponds to the dashed line with an upward slope in Figure 2, and the falling window corresponds to the dashed line with a downward slope in Figure 2. Multiply the T ₀ /4 signal before the last pitch period T ₀ of the history buffer by the rising window, multiply the signal of the last T ₀ /4 of the buffer by the falling window and perform superposition processing, and then replace the last T ₀ of the history buffer /4 signal to ensure a smooth transition where two adjacent pitches meet when the pitch is repeated.

In speech communication, when the discrete cosine transform (DCT) is applied to wideband audio coding, the impact response of the bandpass filter is limited in length, resulting in the so-called block boundary effect, resulting in a lot of noise. The modified discrete cosine transform (Modified DiscreteCosine Transform, MDCT) is usually used to overcome these shortcomings.

MDCT uses Time Domain Aliasing Cancellation (TDAC) technology to reduce "boundary effects". In order to find the MDCT coefficient of a signal composed of 2N samples, first, for the input sequence x[n], use the N samples of this frame and the N samples of the previous adjacent frame signal to form a sequence of 2N samples, and then Define a window function with a length of 2N points as h[n], and h[n] satisfies the following conditions:

h[n] ² +h[n+N] ² = 1(1)

For example h[n] can be simply defined as a sine window:

$\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}}\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

This results in a 50% overlap between the data in each window. The MDCT coefficient X[k] of x[n] and its IMDCT (Inverse Modified Discrete Cosine Transform, improved discrete cosine transform) coefficient Y[n] are defined as:

$\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; [</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; [</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Among them, k=0,..., N-1, n=0,..., 2N-1,

Then the reconstructed signal y[n] can be obtained by performing TDAC on Y[n] and Y′[n] by the following formula:

y[n]=h[n+N]·Y′[n+N]+h[n]·Y[n]n=0, . . . , N-1, (5)

Among them, Y'[n] represents the IMDCT coefficients before and adjacent to Y[n].

At the encoding end, the encoder performs MDCT transformation on the original speech signal according to formula (3) to obtain X[k], and encodes X[k] and sends it to the decoding end; at the decoding end, the decoding end receives the MDCT sent by the encoding end After the coefficients, the decoder performs IMDCT transformation on the received MDCT coefficients X[k] according to formula (4) to obtain the IMDCT coefficients Y[n] corresponding to X[k].

For the convenience of description, it is stipulated that the IMDCT coefficient obtained by the decoder after performing IMDCT transformation on the currently received MDCT coefficient X[k] is Y[n], n=0,...,2N-1, before Y[n] And the adjacent IMDCT coefficients are Y′[n], n=0, . . . , 2N-1. Now take Fig. 3 as an example for illustration. According to the above agreement, the IMDCT coefficient corresponding to frame F0 and frame F1 is IMDCT1, expressed by Y'[n], n=0,...,2N-1; frame F1 and F2 correspond to The IMDCT coefficient of is IMDCT2, represented by Y[n], n=0, . . . , 2N-1. At the decoding end, the decoder puts these two IMDCT coefficients Y[n], n=0,..., 2N-1 and Y'[n], n=0,..., 2N-1, into the formula In (5), the reconstructed signal y[n] can be obtained.

When MDCT coefficients are lost, as shown in Figure 4, the decoder receives the MDCT coefficients MDCT3 corresponding to frames F2 and F3 and the MDCT coefficients MDCT5 corresponding to F4 and F5, but does not receive the MDCT coefficients MDCT5 corresponding to frames F3 and F4. MDCT coefficient MDCT4, then the decoder side can not obtain IMDCT4 according to the formula (4). Therefore, for the decoder, it only receives the coefficients corresponding to F3 of IMDCT3 and the coefficients of F4 corresponding to IMDCT5, and the signals corresponding to frames F3 and F4 cannot be completely recovered by only using IMDCT3 and IMDCT5.

The inventor found in the process of realizing the present invention: the current processing method needs to use the decoded signal before the frame F2 (including F2) to generate the signal of the lost frame, and completely discard the corresponding frame F3 in the received IMDCT3. Part of the coefficients and that part of the coefficients corresponding to frame F4 in IMDCT5. According to the definition of MDCT/IMDCT in formula (3) and formula (4), and in conjunction with formula (5), it can be seen that the corresponding frame F3 coefficient of the received IMDCT3 and the corresponding frame F4 coefficient of IMDCT5 discarded It still contains usable information; moreover, if the frame length is set to L sampling points, then for every n MDCT coefficients lost continuously, the number of sampling points corresponding to the affected signal is (n+1)*L. The more MDCT coefficients are lost, the worse the quality of the recovered signal will be, the worse the user experience will be, and the quality of service (Quality of Service, QoS) will not be guaranteed.

Contents of the invention

The embodiment of the present invention is a method and device for concealing frame loss, which fully utilizes part of the received signals to restore high-quality voice signals and improve QoS.

On the one hand, the embodiment of the present invention proposes a frame loss concealment method, including:

When it is detected that the MDCT coefficient is lost, using the historical signal before the lost frame corresponding to the MDCT coefficient to generate a first composite signal;

Carry out fast IMDCT transformation to described first synthetic signal, obtain the IMDCT coefficient corresponding to described lost MDCT coefficient, IMDCT coefficient calculation formula is as follows;

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

Wherein, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, h[n] represents a window function, x'[n] represents the first composite signal, and N is the frame length;

The IMDCT coefficient corresponding to the lost MDCT coefficient and the IMDCT coefficient adjacent to the IMDCT coefficient are used to perform TDAC processing to obtain a signal corresponding to the lost frame.

On the other hand, an embodiment of the present invention proposes a device for implementing frame loss concealment, including:

A composite signal generation module, configured to generate the first composite signal using the historical signal before the lost frame corresponding to the MDCT coefficient when the loss of the MDCT coefficient is detected;

A fast IMDCT calculation module, configured to perform fast inverse MDCT transformation IMDCT transformation on the first composite signal to obtain the IMDCT coefficients corresponding to the lost MDCT coefficients, and the IMDCT corresponding to the lost MDCT coefficients calculated by the fast IMDCT calculation module The formula for calculating the coefficient is:

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

Wherein, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, h[n] represents a window function, x'[n] represents the first composite signal, and N is the frame length;

The TDAC module is configured to use the IMDCT coefficient corresponding to the lost MDCT coefficient and the IMDCT coefficient adjacent to the IMDCT coefficient to perform TDAC processing to obtain a signal corresponding to the lost frame.

Implementing a method and device for concealing frame loss proposed by the embodiment of the present invention recovers high-quality voice signals and improves QoS by making full use of part of the received signals.

Description of drawings

FIG. 1 is a schematic diagram of a current packet loss concealment technology based on pitch repetition for signal filling;

Figure 2 is a current schematic diagram of smoothing the signal in the pitch buffer

3 is a schematic diagram of the corresponding relationship between MDCT/IMDCT coefficients and signal frames in the prior art;

Fig. 4 is a comparison diagram of signals sent by the encoder and received and decoded by the decoder when packet loss occurs in the prior art;

FIG. 5 is a flow chart of a method for concealing frame loss according to an embodiment of the present invention;

Fig. 6 is the detailed flowchart of step S1 described in Fig. 5;

7 is a schematic diagram of generating a first composite signal based on pitch repetition in an embodiment of the present invention;

8 is a schematic diagram of generating a first composite signal based on pitch repetition in an embodiment of the present invention;

9 is a schematic diagram of generating a first composite signal based on pitch repetition in an embodiment of the present invention;

10 is a schematic diagram of generating a first composite signal based on pitch repetition in an embodiment of the present invention;

FIG. 11 is a schematic structural diagram of a device for implementing packet loss concealment according to an embodiment of the present invention;

FIG. 12 is a schematic structural diagram of the synthetic signal generation module described in FIG. 11 .

Detailed ways

A technical solution of a frame loss concealment method and device proposed in the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

Referring to FIG. 5 , it illustrates a flow chart of a frame loss concealment method according to an embodiment of the present invention. For the convenience of description, in combination with Fig. 4, the decoder receives the MDCT coefficient MDCT3 corresponding to frame F2 and frame F3 and the MDCT coefficient MDCT5 corresponding to F4 and F5, but does not receive the MDCT coefficient MDCT4 corresponding to frame F3 and frame F4. The specific steps of the decoder side processing include:

S1, when the decoder detects that the MDCT coefficient is lost, use the historical signal before the lost frame corresponding to the MDCT coefficient to generate the first composite signal; in this example, the lost frames corresponding to MDCT4 are frame F3 and frame F4, and the historical signal Signals corresponding to the frame F2 and the frame before the frame F2;

S2. Perform fast IMDCT transformation on the first synthesized signal by using a fast inverse transform algorithm to obtain IMDCT coefficients corresponding to the lost MDCT coefficients;

S3. Perform TDAC processing by using the IMDCT coefficient corresponding to the lost MDCT coefficient and the IMDCT coefficient adjacent to the IMDCT coefficient to obtain a signal corresponding to the lost frame corresponding to the lost MDCT coefficient.

During specific implementation, as shown in FIG. 6, in conjunction with FIG. 4 and FIG. 7, in step S1, the generation of the first composite signal using the historical signal before the lost frame corresponding to the MDCT coefficient specifically includes:

S101. Obtain the pitch period T ₀ corresponding to the historical signal before the lost frame;

S102, copying the signal whose last length is T ₀ of the historical signal to the pitch buffer PB ₀ ;

S103. Multiply the historical signal starting from 5T ₀ /4 at the end and having a length of T ₀ /4 by multiplying the signal after a rising window with the signal starting from 3T ₀ /4 in the pitch buffer and having a length of T ₀ /4 The signal is multiplied by a signal after a drop window for cross-fading processing, and the signal obtained by the cross-fading processing is replaced with a signal starting at 3T ₀ /4 in the pitch buffer and having a length of T ₀ /4;

It should be noted that there is no need to update the last T ₀ /4 signal of the historical signal here, because frame F3 still has some effective signals, and the part of the signal near the end of the lost frame is relatively close to the original signal, according to the superposition offset nature, there is no need to cross-fade the tail of the historical signal.

S104, using the signal with length _T0 in the pitch buffer to generate a first composite signal, that is, signals x'[n] corresponding to frames F3 and F4 affected by MDCT4 loss.

Assuming that the signal in the pitch buffer is represented by p ₀ [x], x=0, ..., T ₀ -1, it is synthesized according to formula (6) to obtain x′[n]:

x'[n]=p ₀ [n%T ₀ ], n=0, 1, 2, . . . , 2N-1 (6)

Wherein, N is a non-negative integer, indicating the frame length.

At the same time, the initial value of the initial phase d _offset is 0, and after synthesizing two frames of signals, the phase is updated according to the formula (7):

d _offset = 2N% T ₀ (7)

It should be noted that when MDCT coefficients are continuously lost, the signal x′[n] of the lost frame is synthesized by repeatedly using formula (8):

x'[n]=p ₀ [(n+d _offset )%T ₀ ], n=0, 1, 2, . . . , N-1(8)

After the synthetic signal x′[n] is generated, the phase d _offset is updated according to formula (9):

d _offset = (d _offset + N)% T ₀ , (9)

Among them, N represents the frame length, and d _offset represents the phase.

In the embodiment of the present invention, the step of using the historical signal before the lost frame corresponding to the MDCT coefficient to generate the first composite signal further includes:

Correcting the first synthesized signal by using at least one MDCT coefficient after the lost frame, that is, using the received intact signal after the lost frame to generate x′[n] together to obtain a signal with better quality. The specific implementation method includes two kinds:

Embodiment 1: The first composite signal is corrected by using only one MDCT coefficient after the lost frame:

First adopt the method of step S1 described in Fig. 6 to synthesize the signal x'[n] corresponding to frame F3, frame F4 and frame F5, n=0,...,3N-1; then, perform x'[n] Phase synchronization, as shown in Figure 8. Since only one MDCT coefficient corresponding to the IMDCT coefficient can be used, and the IMDCT coefficient is a damaged signal relative to the original signal, but according to the characteristics of the windowing function, the limited number of sampling points near the junction of the frame F4 and the frame F5 and The amplitude of the original signal is very close, therefore, these limited sampling points can be used to synchronize the phase of the synthesized signal, specifically:

Take the starting sampling point of the IMDCT coefficient corresponding to frame F5 as the midpoint, take M sampling points on both sides as a fixed template window to do waveform matching with the signal x′[n], and use the formula (10) to obtain the phase difference d,

$\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; R</span>}}{\mathrm{<span\; class="notranslate">\; arg</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

Among them, [-R, R] is the allowable range of the phase difference, y'[n], n=0,..., 2N-1 is the IMDCT5 coefficient Y[n], n=0,..., 2N- 1 The lossy signal obtained after windowing according to formula (11), namely:

y'[n]=h[n]·Y[n], n=0,...,2N-1; (11)

Then adjust the synthetic signal according to formula (12), obtain the second synthetic signal x″[n], n=0,...,2N-1:

${\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span><span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; ></span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ></span><span\; class="notranslate">\; =</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; |</span>\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; |</span>\end{array}\right.& \mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

Finally, use the following formula to perform cross-fading on x′[n] and x″[n], and replace x′[n] with the cross-faded signal:

$x^{\&prime;}\left[\mathrm{no}\right]=\frac{2N-\mathrm{no}}{2N+1}\&CenterDot;x^{\&prime;}\left[\mathrm{no}\right]+\frac{\mathrm{no}}{2N+1}\&Center\; Dot;x^{\&prime;\&prime;}\left[\mathrm{no}\right]$ n=0, . . . , 2N-1(13)

In Embodiment 1, a limited number of sampling points are used for phase matching, and if multiple MDCT coefficients are available after a lost frame, the decoded intact signal can be used for phase matching.

Embodiment 2: The first composite signal is corrected by using multiple consecutive MDCT coefficients after the lost frame:

2.1, only for phase synchronization

The method is described in detail below by taking FIG. 9 as an example. Let z[n], n=0, . . . , L-1 be the intact signal after the lost frame, and L be the number of intact sampling points that can be used after the lost frame. As shown in FIG. 9 , z[n], n=0, .

First adopt the method of step S1 described in Fig. 6 to synthesize the signal x'[n] corresponding to frame F3, frame F4 and frame F5, n=0,...,3N-1, then use z[n] to x'[ n] for phase matching to obtain the corresponding phase difference d, specifically: take z[n] to start a signal template with a length of M, and then follow the formula (14) Obtain the phase difference d:

$\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; R</span>}}{\mathrm{<span\; class="notranslate">\; arg</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

Among them, [-R, R] is the allowable range of the phase difference.

After obtaining the phase difference d, then obtain the second composite signal x″[n] according to the formula (12), n=0, ..., 2N-1:

Finally, use the formula (13) to perform cross-fading on the first composite signal x′[n] and the second composite signal x″[n], and replace x′[n] with the cross-faded signal:

2.2, only do backward stacking

In the case of a long frame, the pitch period T ₁ of the current frame signal z[n], n=0, .

In the case of short frames, only using the decoded signal z[n] is not enough to obtain the pitch period T ₁ of the signal corresponding to the current frame, but considering that in the case of short frames, the signal corresponding to the lost frame does not have the pitch period Large changes will occur, so the pitch period T ₀ of the historical signal can be used as the initial value of the pitch period T ₁ corresponding to the current frame, and then fine-tuned on T ₁ to obtain the specific value of T ₁ , specifically:

First initialize T ₁ to the pitch period T ₀ , that is, T ₁ =T ₀ , and then use the Average Magnitude Difference Function (AMDCF) to fine-tune T ₁ to obtain a more accurate T ₁ , specifically, using the formula ( 15) Fine-tuning T ₁ :

${\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="T"\; filenames="S2008100282233E00031.png"\; state="\{\{state\}\}">T</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; arg</span>}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; R</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 15</span>}<span\; class="notranslate">\; )</span>$

Among them, R is the set adjustment range, and it is recommended at 8KHZ sampling rate: R=3

M is the length of the corresponding window when using AMDF. In this embodiment, it is suggested that:

M=min(T ₀ *0.55, LT ₀ ) (16)

z[n] is the intact signal received after the affected frame, and L is the number of available sampling points after the lost frame.

After obtaining T ₁ , copy the first T ₁ samples of z[n] to the pitch buffer PB ₁ to initialize the pitch buffer PB ₁ . The signal in the pitch _buffer PB ₁ is represented _by p ₁ [n], n=0, .

p ₁ [n]=z[n] n=0, . . . , T ₁ -1 (17)

After initializing the pitch buffer PB ₁ , use the backward pitch cycle repetition method to generate the second synthetic signal x''[n], n=0,...,2N-1, specifically:

As shown in Figure 10, frame F2 corresponds to the last good frame before lost frame F3 and lost frame F4, frames F3 and F4 correspond to frames affected by the loss of MDCT coefficients, and frame F5 corresponds to a good frame decoded by the decoder. In the waveform diagram in Figure 10, the signal corresponding to the upper dashed line is the signal x′[n] generated using the historical signal, and the signal corresponding to the lower dashed line is the signal x″[n] generated using the intact signal after the affected frame n].In order to ensure that the speech of the backward pitch cycle repetition filling does not have a waveform mutation at the splicing place of the two pitch cycles, before carrying out the backward pitch cycle repetition filling, the frame F5 needs to be smoothed. The frame F5 is carried out The specific method of smoothing is:

The signal with a starting length of T ₁ /4 in z[n] is multiplied by an ascending triangular window one by one, and the signal with a length of T ₁ /4 at the beginning of a gene cycle length in z[n] is multiplied by a descending triangle The signal behind the window is cross-faded, and the signal with an initial length of T ₁ /4 in the pitch buffer PB ₁ is replaced by the cross-faded signal. Expressed by formula (18) as follows:

$p_1\left[\mathrm{no}\right]=\frac{{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="S2008100282233E00031.png"\; state="\{\{state\}\}">T</figure-callout>}}_1/4-\mathrm{no}}{{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="S2008100282233E00031.png"\; state="\{\{state\}\}">T</figure-callout>}}_1/4+1}*z[{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="S2008100282233E00031.png"\; state="\{\{state\}\}">T</figure-callout>}}_1+\mathrm{no}]+\frac{\mathrm{no}}{{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="S2008100282233E00031.png"\; state="\{\{state\}\}">T</figure-callout>}}_1/4+1}*z\left[\mathrm{no}\right]$ n=0, . . . , T ₁ /4-1 (18)

After frame F5 is smoothed, use the signal of T ₁ sampling points starting from the pitch buffer PB ₁ , and use the pitch repetition method to generate the signal x″[n], as shown by the three arrows in Figure 10, use the formula (19) is expressed as follows:

x″[n]=p ₁ [((T ₁ -2N%T ₁ )+n)%T ₁ ], n=0, Λ, 2N-1 (19)

Finally, x″[n] and x′[n] are cross-faded, and the signal obtained after cross-fading is used to update x′[n] according to formula (13).

It should be noted that when the number M of available sampling points after the lost frame is not enough to meet the smoothing conditions, that is, when T ₁ *1.25<M, only the phase synchronization of the synthesized signal is performed according to the method in method 2.1.

The specific implementation of step S1 has been described in detail above in conjunction with FIGS. 6-10 . The following further describes in the embodiment of the present invention, based on the signal x′[n] obtained in the above implementation, fast IMDCT transformation is performed. Specifically, in the step In S2, according to the properties of MDCT and IMDCT coefficients, the following formula can be used to quickly obtain the IMDCT coefficients corresponding to the lost frame as:

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 20</span>}<span\; class="notranslate">\; )</span>$

Wherein, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, x'[n] represents the first synthesized signal, and N is the frame length.

During specific implementation, the step S3 uses the IMDCT coefficients corresponding to the lost MDCT coefficients and the IMDCT coefficients adjacent to the IMDCT coefficients to perform time-domain aliasing cancellation TDAC processing to obtain the signal corresponding to the lost frame Specifically include:

Perform superposition processing according to formula (5) to obtain the signal corresponding to the lost frame;

Wherein, y[n] represents the signal corresponding to a lost frame corresponding to the lost MDCT coefficient, h[n] represents the window function used for TDAC processing, and Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, Then Y'[n+N] represents the IMDCT coefficients adjacent to Y[n] and before Y[n].

In the embodiment of the present invention, the first N coefficients of IMDCT4 obtained in step S2 and the last N coefficients of INMDCT3 are superimposed, and the signal y ₁ [n] corresponding to frame F3 is obtained as follows:

y ₁ [n]=h[n+N]·Y ₁ '[n+N]+h[n]·Y ₁ [n]n=0, . . . , N-1,

Y ₁ [n]=h[n] x'[n]-h[Nn-1] x'[Nn-1]n=0,..,N-1;

Among them, Y ₁ [n] represents the IMDCT coefficient corresponding to frame F3 (i.e. the first N coefficients of IMDCT4), Y ₁ '[n+N] represents the IMDCT coefficient corresponding to the frame F2 (i.e. the last N coefficients of IMDCT3) , N represents the frame length;

Utilize the last N coefficients of IMDCT4 obtained in step S2 and the first N coefficients of IMDCT5 to superimpose, obtain the signal y ₂ [n] of frame F4 as:

_y2 [n]=h[n+N]· _Y2 '[n+N]+h[n]· _Y2 [n]n=N, . . . , 2N-1,

Y ₂ [n]=h[n] x'[n]-h[3N-n-1] x'[3N-n-1]n=N, ..2N-1;

Among them, Y ₂ [n] represents the IMDCT coefficient corresponding to frame F4 (that is, the last N coefficients of IMDCT4), Y ₂ ′[n+N] represents the IMDCT coefficient corresponding to frame F5 (the first N coefficients of IMDCT5), and N represents frame length.

As can be seen from the above, by implementing a method for concealing lost frames in the embodiment of the present invention, the signal of the lost frame is recovered by using the partial signal of the lost frame and the intact signal after the lost frame, which makes full use of the signal resource and improves the user's performance. Experience, guaranteed QoS.

A device for implementing frame loss concealment according to an embodiment of the present invention will be described in detail below with reference to FIG. 11 and FIG. 12 .

As shown in Figure 11, a device for realizing frame loss concealment includes:

The composite signal generating module 100 is configured to synthesize the first composite signal using the historical signal before the lost frame corresponding to the MDCT coefficient when the loss of the MDCT coefficient is detected;

The fast IMDCT calculation module 200 is configured to use a fast inverse transform algorithm to perform fast inverse MDCT transform IMDCT transform on the first composite signal to obtain IMDCT coefficients corresponding to the lost MDCT coefficients;

The TDAC module 300 is configured to use the IMDCT coefficients corresponding to the lost MDCT coefficients and IMDCT coefficients adjacent to the IMDCT coefficients to perform TDAC processing to obtain signals corresponding to the lost frames.

During specific implementation, as shown in FIG. 12, the synthetic signal generation module 100 includes:

An acquisition unit 101, configured to acquire a historical signal before the lost frame and a pitch period corresponding to the historical signal;

The copying unit 102 is used to copy the last signal whose length is the pitch period of the historical signal acquired by the acquiring unit to the pitch buffer;

a pitch buffer unit 103, configured to buffer a signal of a pitch cycle length copied by the copying unit;

The cross-fading unit 104 is used to multiply the signal after the last 5T ₀ /4 of the historical signal with a length of T ₀ /4 by the signal after a rising window and start at 3T ₀ /4 in the pitch buffer, and the length is The signal of T ₀ /4 is multiplied by the signal after a drop window for cross-fading processing, and the signal obtained by the cross-fading processing is replaced with the signal starting at 3T ₀ /4 in the pitch buffer and having a length of T ₀ /4 , where T ₀ represents the pitch period;

The synthesis processing unit 105 is configured to generate a first synthesized signal by using a pitch repetition method according to the signal with a length _T0 in the pitch buffer.

Wherein, the first synthetic signal is:

x'[n]=p ₀ [n%T ₀ ], n=0, 1, 2, . . . , 2N-1,

Wherein, p ₀ [x], x=0, . . . , T ₀ -1 represents the signal in the pitch buffer, wherein T ₀ represents the pitch period, and N represents the frame length.

When detecting continuous loss of MDCT coefficients, the first composite signal is:

x'[n]=p ₀ [(n+d _offset )% T ₀ ], n=0, 1, 2, . . . , N-1,

d _offset = (d _offset + N)% T ₀ ,

Wherein, T ₀ represents the pitch period, N represents the frame length, d _offset represents the phase, and its initial value is 0.

During specific implementation, the synthetic signal generating module 100 further includes:

The correction unit 106 is configured to use at least one MDCT coefficient after the lost frame to correct the first synthesized signal synthesized by the synthesis processing unit 105, specifically including two implementation modes: using only one MDCT coefficient after the lost frame Perform correction; perform correction by using multiple consecutive MDCT coefficients after the lost frame, which has been described in detail above in conjunction with FIGS. 8-10 , and will not be repeated here.

During specific implementation, the fast IMDCT calculation module 200 uses a fast inverse transform algorithm to perform fast inverse MDCT transform IMDCT transform on the first composite signal, and obtain the IMDCT coefficients corresponding to the lost MDCT coefficients as:

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

x'[n] represents the first composite signal, and N is the frame length.

During specific implementation, the TDAC module 300 uses the IMDCT coefficients corresponding to the lost MDCT coefficients and the IMDCT coefficients adjacent to the IMDCT coefficients to perform TDAC processing, and obtains the signal corresponding to the lost frame corresponding to the lost MDCT coefficients as :

y[n]=h[n+N]·Y′[n+N]+h[n]·Y[n]n=0,...,N-1

Among them, h[n] represents the window function used for TDAC processing, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, then Y'[n+N] represents the previous one adjacent to Y[n] IMDCT coefficients.

It should be noted that those of ordinary skill in the art should understand that a frame loss concealment method proposed in the embodiment of the present invention can be realized by programming a computer program, an instruction, or a programmable logic device, and the program can be saved to a memory On media, such as: CD, disk, etc.

It can be seen from the above-mentioned embodiments that according to a method and device for concealing frame loss according to the embodiments of the present invention, in the superposition mode, according to the properties of MDCT, the IMDCT coefficients of the synthesized signal are obtained by using a low-complexity fast algorithm, and by making full use of the received part of the signal, restore high-quality voice signals, and improve QoS.

The above description is a preferred embodiment of the present invention, and it should be pointed out that for those skilled in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications are also considered Be the protection scope of the present invention.

Claims (6)

1. A method for frame loss concealment, comprising: When the loss of the improved discrete cosine transform MDCT coefficient is detected, the first composite signal is generated by using the historical signal before the lost frame corresponding to the MDCT coefficient; Inverse discrete cosine transform (IMDCT) transformation is carried out to described first synthesis signal rapidly, obtains the IMDCT coefficient corresponding to the MDCT coefficient of described loss, and the calculation formula of IMDCT coefficient is as follows; $\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$ Wherein, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, h[n] represents a window function, x'[n] represents the first composite signal, and N is the frame length; Using the IMDCT coefficients corresponding to the lost MDCT coefficients and IMDCT coefficients adjacent to the IMDCT coefficients to perform time-domain aliasing cancellation TDAC processing to obtain signals corresponding to the lost frames. 2. The method according to claim 1, wherein generating the first synthetic signal using the historical signal before the lost frame corresponding to the MDCT coefficient comprises: Obtaining a historical signal before the lost frame and a pitch period corresponding to the historical signal; Copying the last signal whose length is T ₀ of the historical signal to the pitch buffer, where T ₀ represents the pitch period; Multiply the signal obtained by multiplying the signal with a length of T ₀ /4 starting at 5T ₀ /4 from the last 5T 0 /4 of the historical signal by a rising window and the signal starting at 3T ₀ /4 in the pitch buffer and having a length of T ₀ /4 The signal obtained after being multiplied by a drop window is subjected to cross-fading processing, and the signal obtained by the cross-fading processing is replaced with a signal starting at 3T ₀ /4 in the pitch buffer and having a length of T ₀ /4; Using the signal with length _T0 in the pitch buffer to generate the first synthesized signal by using pitch repetition. 3. The method according to claim 2, wherein the step of generating the first synthetic signal by using the historical signal before the lost frame corresponding to the MDCT coefficient further comprises: The synthesized first synthesized signal is corrected using at least one MDCT coefficient following the lost frame. 4. A device for realizing frame loss concealment, comprising: A composite signal generating module, configured to generate a first composite signal using the historical signal before the lost frame corresponding to the lost MDCT coefficient when the loss of the MDCT coefficient is detected; The fast IMDCT calculation module is used to perform fast IMDCT transformation on the first composite signal to obtain the IMDCT coefficients corresponding to the lost MDCT coefficients, and calculate the IMDCT coefficients corresponding to the lost MDCT coefficients calculated by the fast IMDCT calculation module The formula is: $\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span>& \mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$ Wherein, Y[n] represents the IMDCT coefficient corresponding to the lost MDCT coefficient, h[n] represents a window function, x'[n] represents the first composite signal, and N is the frame length; The TDAC module is configured to use the IMDCT coefficients calculated by the fast IMDCT calculation module and the IMDCT coefficients adjacent to the IMDCT coefficients to perform TDAC processing to obtain the signal corresponding to the lost frame. 5. The device according to claim 4, wherein the synthetic signal generation module comprises: an acquisition unit, configured to acquire a historical signal before the lost frame and a pitch period corresponding to the historical signal; The copy unit is used to copy the last signal whose length is the pitch period of the historical signal acquired by the acquisition unit to the pitch buffer; a pitch buffer unit, configured to buffer a signal of a pitch cycle length copied by the copy unit; The cross-fading unit is used to multiply the signal with a length of T _{0 /4 starting from 5T 0 /} 4 of the historical signal with the signal after a rising window and the signal starting from 3T ₀ /4 in the pitch buffer and having a length of T The signal of _0/4 is multiplied by the signal after a drop window for cross-fading processing, and the signal obtained by the cross-fading processing is replaced with the signal starting at 3T _0/4 in the pitch buffer and having a length of T _0/4 , Wherein T ₀ represents the pitch period; A synthesis processing unit, configured to generate a first synthesis signal by using a pitch repetition method according to the signal with a length _T0 in the pitch buffer. 6. The device according to claim 5, wherein the synthetic signal generation module further comprises: A correction unit, configured to use at least one MDCT coefficient after the lost frame to correct the first composite signal synthesized by the synthesis processing unit.

Images